Proc. Nati. Acad. Sci. USAVol. 82, pp. 1030-1034, February 1985Biochemistry

Bovine papillomavirus contains multiple transforming genes(transcription/cDNA cloning/expression vector/papovavirus/transformation)

YU-CHUNG YANG*, HIROTO OKAYAMAt, AND PETER M. HOWLEY**Laboratory of Pathology, National Cancer Institute, Bethesda, MD 20205; and Laboratory of Molecular Genetics, National Institute of Child Health andDevelopment, Bethesda, MD 20205

Communicated by DeWitt Stetten, Jr., October 16, 1984

ABSTRACT Bovine papillomavirus type 1 (BPV-1) and itscloned full-length DNA can transform rodent cells in vitro, andthe viral DNA persists as an extrachromosomal multicopyplasmid in these transformed cells. Previous studies have iden-tified at least five discrete viral RNAs that are expressed inBPV-1 transformed cells and have shown that these transcriptsshare a 3' coterminus. To further define the structure of theseRNAs and to characterize the functions of individual viraltranscripts, we constructed a cDNA library with mRNA fromBPV-1-transformed mouse C127 cells using an Okayama andBerg plasmid. From a library of 105 independent clones, 200BPV-1 specific clones were isolated and characterized. Se-quence analysis has revealed differential splicing patterns forthe mRNA species in BPV-1 transformed cells. In conjunctionwith the open reading frames (ORFs) deduced from the BPV-1DNA sequence, it is possible to predict the structure of thepotential encoded proteins. The vector used to generate thesecDNA clones contains mammalian cell transcriptional regula-tory elements, facilitating their functional characterization.We have identified two distinct classes of cDNA clones that caneach independently transform mouse C127 cells. One class ofcDNA clones contains the- E2 ORF intact and the second con-tains the E6 ORF intact. These two putative viral functionsappear to act synergistically in transforming mouse C127 cellsin vitro.

fied several regions of BPV-1 genome that influence theexpression of the viral transforming functions (9-12). Dele-tion mutants affecting the E2 ORF are significantly impairedin their ability to transform mouse C127 cells, suggesting thatthe putative E2 protein is an important transforming protein(9). The expression of E2 alone, however, is not sufficientfor the fully transformed phenotype. Deletion mutagenesisstudies map an additional function that is critical for the fullytransformed phenotype to the region between the Hpa I site(base 1) and the Sma I site (base 945) (9). These previousstudies did not distinguish whether these two distinct regionsof the viral genome encode separate proteins or whetherthey contain exons that are spliced together at the level ofmRNA to generate a single transforming protein.To further define the structure of the mRNAs in trans-

formed cells and to characterize the functions encoded byindividual viral transcripts, we constructed a cDNA libraryusing mRNA from BPV-1-transformed mouse cells. The vec-tor used for cDNA cloning contains the simian virus 40(SV40) early promoter, the SV40 late region introns, and theSV40 late region polyadenylylation site, allowing expressionof cDNA clones in mouse cells for functional analysis (17).In this study, expression vectors that carried individualcDNA inserts were tested for their ability to transformmouse C127 cells.

Bovine papillomavirus type 1 (BPV-1) causes benign fibro-papillomas (warts) in cattle and also induces fibrosarcomasin heterologous hosts. The virus or its molecularly clonedDNA can transform certain mouse cells in vitro, and theDNA persists as an extrachromosomal multicopy plasmid inthe transformed cells (1). A 69% subgenomic fragment(BPV69T) bounded by the unique HindIII and BamHI sites issufficient to induce transformation in vitro (2), and recombi-nants containing this DNA segment can be maintained asfree plasmids in transformed cells (3). Cell lines establishedfrom transformed foci have a fully transformed phenotype inthat they are anchorage independent and are tumorigenic innude mice (4).The genomic organization of BPV-1 has been established

from the DNA sequence (5) and from the transcription datafrom both BPV-1-transformed cells (6) and productively in-fected bovine fibropapillomas (7, 8). All of the detectablemRNA species are transcribed from a single strand, and allof the open reading frames (ORFs) of >400 base pairs arelocated on the same strand. As indicated in Fig. 1, severalmRNAs have been mapped in the 69% transforming regionwhere eight of the ORFs (labeled El through E8) are located.These viral transcripts have a common 3' end at 0.53 mapunits, but they have bodies that appear to vary at their 5'ends (6).Recent studies from our laboratory and others have identi-

MATERIALS AND METHODSCells and Transformation. Mouse C127 cells (4) and BPV-

1-transformed mouse C127 cells (ID13) (1) were maintainedin Dulbecco's modified Eagle's medium supplemented with10% fetal calf serum. DNA transformation was carried out asdescribed (14).

Construction of cDNA Library with pCD Expression Vec-tor. Polyadenylylated mRNA from ID13 cells was preparedby methods previously described (15, 16). A cDNA librarywas constructed from this mRNA with the plasmids pcDV1and pL1, using the plasmid primer method described byOkayama et al. (17). The bacterial transformation was car-ried out using Escherichia coli strain DH1 (13).

Isolation and Characterization of BPV-1 cDNA Clones.BPV-1 specific cDNA clones were identified by colony hy-bridization (18, 19). Viral cDNA clones with inserts >1 kilo-base (kb) were analyzed by restriction enzyme mapping andwere sequenced by Maxam and Gilbert techniques (20).

Analysis of Cellular DNA. Total cellular DNA was extract-ed from cDNA transformed cells according to a modifiedmethod described previously (1). Southern blot analysis wasperformed as described (21).

RESULTSScreening of cDNA Library for BPV-1 Sequences. A library

of -100,000 independent clones was made from the polya-

Abbreviations: BPV, bovine papillomavirus; ORF, open readingframe; SV40, simian virus 40; kb, kilobase(s).

1030

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 82 (1985) 1031

5, 3'

1 149 501 1099 1479

2

449 859

CD-Ur 2_-Q=S U;;

LZZ]2581 3837

3173 3526 4172

2663 3267 3551 3714 4010

&-- EU~c

I I I I ...I.a.I

co(V) LOr

r%.- 0 -C) - t .(IN CN CN

FIG. 1. Genomic organization of BPV-1 DNA. The full-length molecule (7945 base pairs) of BPV-1 opened at the unique HindI1 (base 6959)is marked off with restriction sites and bases noted at the bottom of this figure. The transforming segment from HindIII to the BamHI site isindicated by the heavy bar (2). The region transcribed in transformed cells and the direction of transcription are indicated by the arrow at the topof the figure (6). Open bars represent potential coding regions for BPV-1 proteins in each of the three reading frames (5). ORFs within thetransforming region have been designated E1-E8. Numbers beneath ORF represent the first and last bases of the ORF. This schematicrepresentation of the BPV-1 genome has been published previously (9) and is reproduced with permission.

denylylated RNA from ID13 cells. Using a BPV-1 specificradiolabeled probe, the library was screened and 200 BPV-1-specific clones were isolated. Positive clones were then hy-bridized with subgenomic fragments of BPV-1 DNA to pro-vide preliminary mapping data and information on which todesign a sequencing strategy. The insert sizes of viral cDNAclones ranged from 0.2 to 2.8 kb. Since the viral RNAs inBPV-1-transformed cells range in size from -1 to 4 kb (6),only clones that contained inserts -1 kb or larger were ana-lyzed further.

Structural Analysis of BPV-1 Specific cDNA Clones. Thedata summarizing the cDNA clones that were sequenced arepresented in Table 1. Seven classes of cDNA clones can bedistinguished based on these sequence data. Previous analy-sis of the RNAs present in BPV-1-transformed cells has indi-cated that the different viral transcripts have a 3' coterminusat '0.53 map units. Sequence analysis of several cDNAslisted in Table 1 indicates that the mRNA is polyadenylylat-ed at base 4203, 24 bases 3' to the first A of the polyadenylyl-ation recognition site A-A-T-A-A-A.

In five of the seven classes of cDNAs characterized,spliced messages were identified. Splice donor sites wereidentified at bases 304, 864, and 2505. Splice acceptor siteswere found at bases 527 and 3224. The DNA sequence ofeach of these splice donor and acceptor sites is presented inFig. 2 and agrees reasonably well with the consensus splicedonor and acceptor sequence (22). The 5' end of three of thecDNAs sequenced was at base 89. This agrees with the dataof Ahola et al. (23), who mapped this as a major 5' end of themRNAs present in transformed cells. There is a T-A-T-A-A-A sequence 31 bases upstream at base 58. This sequence hasbeen shown in many instances to be part of an RNA poly-merase II transcriptional promoter and to apparently func-tion in positioning the 5' ends of the mRNAs (24). The classIII cDNA clone (C58) has a 5' end that maps upstream fromthis promoter element, indicating that promoter sequencesother than those with the "TATA" box at nucleotide 58 func-tion in BPV-1-transformed cells. Many of the cDNA cloneshave 5' ends mapping downstream from base 89. Whereassome of these may represent copies of truncated viral RNAs,some could be representative of full-length RNAs possiblygenerated from an additional viral promoter. In the case ofclass V cDNAs and one of the class VI cDNAs (C88) with 5'ends mapping to base 2440, the sequence T-A-A-T-A-T-T is

located 27 nucleotides upstream.

We also identified several unspliced cDNA clones anddesignated them as class VI or class VII clones based onsize. The class VI cDNA clones ranged from 1.9 to 2.5 kband had 5' ends mapping from base 1889 to base 2534. Theclass VII cDNA clones measured from 1.1 to 1.2 kb and had5' ends from base 3212 to base 3309.

Potential Coding Regions for Individual cDNA Clones. Po-tential coding sequences for six of the seven classes ofcDNAs are presented in Fig. 3. The class VII clones are notincluded because they are likely to consist of cDNA copiesof truncated viral mRNAs. Since the viral proteins producedin BPV-1-transformed cells have not yet been identified, thelikely proteins encoded by each of the cDNA classes were

Table 1. Structure of various cDNA clones

Donor (D) andSize of Intron acceptor (A)

Class Clone cDNA, kb 5' end (±) sites

I C102 1.7 89 + 304(D), 527(A)864(D), 3224(A)

II C119 1.4 89 + 304(D), 3224(A)C211 1.4 89 + 304(D), 3224(A)

III C58 2.1 7879 + 864(D), 3224(A)IV C48 1.6 429 + 864(D), 3224(A)

C91 1.4 670 + 864(D), 3224(A)V* C5 1.3 2440 + 2505(D), 3224(A)

C14 1.3 2440 + 2505(D), 3224(A)C37 1.3 2440 + 2505(D), 3224(A)C174 1.3 2440 + 2505(D), 3224(A)C189 1.3 2440 + 2505(D), 3224(A)

VI C52 2.5 1889 -C59 2.1 2360 -C61 2.1 2360 -C88 2.0 2440 -C212 1.9 2534 -

VIIt C36 1.2 3212 -C122 1.2 3233 -C79 1.1 3245 -C32 1.1 3309 -

*Other similar or identical clones include C9, C13, C18, C23, C134,C153, C176, C180, and C207.tOther similar or identical clones include C125, C97, C114, C117,and C199.

Open ReadingFrames:

3 EL813

_

._ ._

II

am

L2 15593/5597

cnI

0) CV)

co r%-

Li I7093

I)

0Y)

2 .nx I

I uiCV

0))CD CD

Biochemistry: Yang et aL

i r. ;oCV) m LO

A r- ODc CV) CV) 3

Proc. Natl. Acad. Sci USA 82 (1985)

Base no Viral donor site Base no Viral acceptor site

304 CAGGTGAGG 527 TCACCTGCAGGA

864 AAGGTAGCA 3224 CTGATTTTAGAG

| 2505 CGGGTGAGC I

EXON * INTRON g - EXON

A 4 AAGGT AGT

C G

C JGPyPyPyPyPyPyX AGG

T T

Consensus Sequences

predicted based on the general rules for protein translation ineukaryotic cells (25).The class I cDNA clone (C102) could direct the synthesis

of a 21-kDa protein that is derived from the in-frame splicingof the E6 and E7 exons, and uses bases 91-93 (AUG) and860-862 (UAA) for the initiation and termination codons.C102 contains a second splice from base 864 to base 3224;however, these sequences are located 3' to the terminationcodon that would be used. The class II clone (C119) splicesthe E6 and E4 ORFs together and would direct the synthesisof a 20-kDa protein using the initiation codon AUG at bases91-93 and the termination codon UGA at bases 3527-3529.The class III cDNA (C58) has a splice donor at base 864 afterthe E6 termination codon UAG at bases 502-504 and, there-fore, would direct the synthesis of the 15.5-kDa E6 protein.The class IV cDNA (C48) with a 5' end at base 429 could be atruncated class III cDNA. If, however, it is representative ofa class of an mRNA species in BPV-1-transformed cells, itcould direct the synthesis of a 15.5-kDa E7 protein with an

FIG. 2. Comparison of BPV-1 splice junctionswith consensus acceptor and donor sites.

AUG initiation codon at bases 479-481 and the terminationcodon at bases 860-862.A number of 1.3-kb clones were identical in structure and

are represented by the class V cDNA clones. These each hada 5' end at base 2440 and were spliced from base 2505 to base3224. These do not contain a large open translation framefollowing an initiation methionine codon. The sequence ofeach of these cDNAs contains stretches of adenines aroundbase 2440 instead of the BPV-1 genomic sequence G-A-T-A-T-T-T located at base 2440. It is possible that the cDNAsrepresented by the class V structures are faithful copies ofRNAs within cells that have been modified in the sequencesat their 5' ends as the result of some specific RNA degrada-tion or cleavage process. Pettersson and his group have ob-tained evidence by heteroduplex analysis of a class of RNAmolecules in BPV-1-transformed cells with 5' ends that mapto this region (Ulf Pettersson, personal communication).Alternatively, these molecules might have been generated bythe staggering effect of reverse transcriptase in a region of

1 89 (C1 ) E6/E7304 527 864 3224 6/

304 89 t)E6/E43224

111 7879 324(C58) E6V864 )73224

IV 429 Az - - - - -(C48) E7864

V

VI

c ca EI (n

2440 Cz-E1 E1/E42505 3224

1889 =. (C52 E2

23605 (C9E2

2440{C88) E2

2534 i IjC212) E2

c == EQ zz(,z co c coi HLn m Lo.q r- oocn c( cn)

_ Fr C%4

Q °0 Qr

CL CJ Um n-1) wiM 1

LC) COCCY) LOlr%.q* *- . *

'r. C14 C4 "

FIG. 3. Structure of various cDNA clones are indicated with the most 5' coding region following an AUG indicated by open box except forthe class V clone, which is believed to be a truncated E1/E4 clone. The putative nontranslated regions are indicated by solid line and interven-ing sequences are shown by slanted lines. Numbers outside the coding region represent 5' ends and splice junctions. Potential ORFs expressedfor each individual cDNA clone are designated on the right. A restriction map is at the bottom of the figure for orientation.

a0

r-

m cx iri iC') a)C') LOI- a)(D (0

1032 Biochemistry: Yang et aL

Proc. NatL. Acad. Sci. USA 82 (1985) 1033

high secondary or tertiary structure of the mRNA. Thesplice from base 2505 to base 3224 results in the fusion of theEl and E4 ORFs; however, there is no methionine initiationcodon in this fused open segment. Whether class V cDNAclones represent truncated copies of transcripts that exist inBPV-1-transformed cells for the production of a protein re-sulting from fusion of the El and E4 ORFs remains to beexamined.The five cDNA clones represented by class VI were un-

spliced and contained the E2 ORF intact; the first AUG islocated at bases 2608-2610 within the E2 ORF. Each of theclass VI cDNAs could, therefore, direct the synthesis of a48-kDa E2 protein with initiation and termination codons atbases 2608-2610 and 3838-3840, respectively. These cDNAsalso contain the E3, E4, and E5 ORFs intact. However, theE3 and E4 ORFs overlap the E2 ORF and are located 3' tothe methionine initiator codon for E2, and, therefore, wouldbe unlikely to be translated from the mRNAs represented bythis class of cDNAs (25). The first AUG in the E5 ORF is atnucleotides 3879-3881. This is downstream from the E2ORF termination codon and thus this region could potential-ly be translated into a 44 amino acid protein from any of thecDNAs in this study if they represented bicistronic mes-sages.The class VII cDNA clones are also unspliced and most

likely represent cDNA copies of truncated mRNAs. ThesecDNA clones contain the COOH end of the E2 ORF and atruncated E4 ORF, as well as the E3 and E5 ORFs. Otherthan the short E5 ORF described above, there are no largeopen reading frame segments in these cDNA clones preced-ed by a methionine AUG.

Transformation of Mouse C127 Cells by Various cDNAClones. Since each of the cDNAs was cloned into an expres-sion vector containing SV40 transcriptional regulatory ele-ments, it was possible to assay biological functions directly.The transforming ability of individual cDNA clones fromeach of the class I to class VI was tested by focus assay onmouse C127 cells, and the results are summarized in Table 2.The cloned cDNAs from classes III and VI were each able totransform mouse C127 cells. The class VI cDNA clones arenonspliced and contain the full E2 ORF, which would mostlikely be translated as discussed above. The class III clone(C58) is spliced and contains the complete E6 ORF as its first

Table 2. Transformation of mouse C127 cells by BPV-1cDNA clones

Potential Foci percoding Foci per plate with

Clone Class region* plate alonet C58 DNAt

C52 VI E2 27, 36, 50 65, 105, 112C59 VI E2 20, 56, 58 82, 98, 102C88 VI E2 7,8, 14 27,32, 66C212 VI E2 7, 13, 18 17, 41, 49C58 III E6 3,4 -C48 IV E7 0,0 -C102 I E6/E7 0, 0 -C119 II E6/E4 0, 0C5 V Truncated 0, 0

E1/E4p142-6 Full BPV-1 100, 120, 136 -

DNASalmon 0, 0spermDNA

*Each of these classes of cDNAs also contains intact E3 and E5ORFs; in addition, class VI contains an intact E4.

major open translation frame behind an initiator methioninecodon.Each of the cDNA clones tested transformed mouse cells

at an efficiency less than that of the full BPV-1 genome (Ta-ble 2). There was some variability among the four class VIclones tested (Table 2), which differ by the length of theBPV-1 segment 5' to the AUG codon in the E2 ORF. Theclass III cDNA clone (C58) induced foci at an efficiency of3-4 foci per plate compared to that of the full BPV-1genome, which induced 100-136 foci per plate in this experi-ment. The size and morphologic appearance of the foci in-duced by each of these cDNAs expressed behind the SV40early promoter are similar. Individual foci induced by theclass III cDNA (C58) and the class VI cDNA (C88) were

cloned into lines and shown to contain between 10 and 20copies of the cDNA vector integrated (data not shown).To test whether transforming products encoded by the

class III and class VI cDNA clones might have a cooperativeeffect in transforming mouse C127 cells, cotransformationexperiments were performed. As shown in Table 2, whenthese cDNA clones were cotransfected, the frequency of fo-cus formation was higher than transfection with either theclass III or the class VI clones alone. This effect is more thanadditive, suggesting a potential cooperative effect betweenthe putative functions encoded by these two classes ofcDNA clones. In addition, the foci induced by cotransforma-tion were larger and appeared faster than those induced byeither the class III or the class VI cDNA clones alone.

DISCUSSIONTranscriptional studies of the papillomaviruses have beenlimited because of the low abundance of viral mRNA intransformed cells (6, 23). The structural analysis of thecDNAs cloned from BPV-1-transformed cells presented inthis present study indicates that the viral RNAs are generat-ed by differential splicing. Splice donor sites map to nucleo-tides 304, 864, and 2505; and splice acceptor sites map tonucleotides 527 and 3224. The different classes of RNA de-tected and their splice junctions are in good agreement withthe studies of Pettersson and colleagues who have indepen-dently mapped the polyadenylylated viral RNA species pres-

ent in BPV-1-transformed cells by heteroduplex analysis (U.Pettersson, personal communication).We map a major 5' end to nucleotide 89, confirming the

results of Ahola et al. (23). A T-A-T-A-A-A sequence at nu-

cleotide 58 is likely an element of a transcriptional promoterlocated upstream from this transcriptional start site. Thepresence of -165 base pairs of sequences upstream frombase 89 in one cDNA clone sequenced indicates that there isan upstream start site(s) and that there is an additional pro-moter(s) functional in BPV-1-transformed cells. Many of the5' ends of the cDNAs analyzed were located downstreamfrom nucleotide 89 and many mapped to approximately nu-

cleotide 2440. Whereas some of these may represent copiesof truncated RNAs or incomplete copies of RNAs, othersmay be true 5' ends, indicating that additional RNA poly-merase II promoter elements may be located in this region.More detailed studies of BPV-1 transcription in transformedcells will be required to answer this question. Prevous stud-ies have indicated that the viral mRNAs have a 3' coter-minus at 0.53 map units. Sequence analysis indicates that thesite of polyadenylylation is at the guanine residue at base4203, located 24 bases downstream from the polyadenylyla-tion recognition sequence at 4180 (A-A-U-A-A-A). The 3'ends ofRNA transcripts in mammalian cells are polyadenyl-ylated at sites 10-25 nucleotides 3' to the A-A-U-A-A-A sig-nal (26).

Previous studies have localized a transforming gene forBPV-1 to the 2.3-kb segment of the BPV-1 genome between

tEach plate received a total of 5 ,ug of the indicated cDNA clone.tEach plate received a total of 5 ,ug of the indicated cDNA clone and5 Ag of C58 DNA.

Biochemistry: Yang et aL

Proc. Natl. Acad. Sci. USA 82 (1985)

the EcoRI site and the BamHI site (9, 11). Furthermore, de-letion mutagenesis studies and targeted base mutagenesisstudies implicate the product of the E2 ORF as having animportant role in BPV-1-mediated cellular transformation(ref. 9; D. DiMaio, personal communication; M. Lusky andM. Botchan, personal communication).Although the class VI cDNA clones that contain the E2

ORF intact could represent true copies of mRNAs thatwould direct the synthesis of a 48-kDa E2 protein in trans-formed cells, they are not spliced and their 5' ends map in aregion where no viral transcriptional promoter has yet beenlocalized. Thus, the true structure of the putative E2 ORFmRNA remains uncertain. Each of the cDNAs in this classis, however, able to independently transform mouse cellswith the expression vectors used in this study, further impli-cating E2 as important in papillomavirus-induced transfor-mation.Our previous studies have mapped a function critical for

the expression of the full transformed phenotype to the seg-ment of the genome containing the E6 and E7 ORFs (9). Thepresent study provides the structure of a number of cDNAswith coding exons located in this region. Of these, only theclone that would direct the synthesis of the E6 ORF intact iscapable of transforming C127 cells, indicating that the 15.5-kDa product of the E6 ORF is a second and independenttransforming protein of BPV-1. A similar conclusion hasbeen independently made by Schiller et al. (27).The mechanism by which the putative products of the E2

and E6 ORFs may lead to transformation is not clear at thispoint. Comparison of amino acid sequences deduced fromdifferent papillomavirus DNA sequences has provided someinteresting characteristics of these proteins. The amino ter-minus of the E2 protein is well conserved among BPV-1,HPV-la, HPV-6, and CRPV (28, 29). There is also aminoacid and structural homology involving "80 amino acids atthe carboxyl terminus of the E2 ORFs of these viral ge-nomes. In addition, this carboxyl-terminal domain has limit-ed homology with the human c-mos gene product (30). Thereis an intriguing property of E6 ORF shared by BPV-1 and theother papillomaviruses sequenced (28, 29). Although thisORF is not well conserved among these sequenced genomes,there is a tetrapeptide sequence, Cys-X-X-Cys, that is re-peated four times in the E6 ORF of each genome with con-served distances between the repeating units. Cysteine-richrepeated segments are also found in the small t antigens ofthe polyomaviruses including SV40, polyoma, and BK. Anumber of studies have indicated a role for SV40 small t anti-gen in transformation and anchorage-independent growth(31, 32).

In this paper, we have established that two classes ofcDNA clones (class III and class VI) can independentlytransform C127 cells, and that they apparently act synergisti-cally in vitro in a focus assay. Although other classes ofcDNAs were negative by focus assay, our studies do not ruleout a potential role for any of the putative viral gene prod-ucts encoded by these clones in BPV-1 transformation. Fur-thermore, additional viral gene products not necessarily re-flected by the cDNAs described in this study may also beinvolved in cellular transformation and may even be able toinduce foci independently.

It seems likely that the transforming capability of the classIII and class VI cDNA clones is due to the different viralproteins they encode. These classes of cDNAs could encodethe putative E6 and E2 proteins, respectively, and deletionstudies have mapped functions involved in transformation tothe regions of the BPV-1 genome containing these ORFs (9).Our interpretation that the transforming functions are due tothese putative proteins, however, must be tempered because

of our current inability to directly define the viral proteinspresent in cells transformed by each of these classes ofcDNAs. A potential role of the downstream ORFs present inthese cDNA clones (i.e., E3, E4, or E5) cannot be ruled outat this time. Deletion mutagenesis studies using the cDNAmolecules described in this paper should permit a resolutionof this issue.

We are grateful to Steve Lussos for technical assistance; to PaulDoherty and Carl Baker for comments and advice; to Carl Baker,George Khoury, and C. Richard Schlegel for critical readings of thismanuscript; and to Susan Hostler for her editorial assistance.

1. Law, M.-F., Lowy, D. R., Dvoretzky, I. & Howley, P. M.(1981) Proc. Natl. Acad. Sci. USA 78, 2727-2731.

2. Lowy, D. R., Dvoretzky, I., Shober, R., Law, M.-F., Engel,L. & Howley, P. M. (1980) Nature (London) 287, 72-74.

3. Sarver, N., Gruss, P., Law, M.-F., Khoury, G. & Howley,P. M. (1981) Mol. Cell. Biol. 1, 486-496.

4. Dvoretzky, I., Shober, R. & Lowy, D. R. (1980) Virology 103,369-375.

5. Chen, E. Y., Howley, P. M., Levinson, A. D. & Seeburg,P. H. (1982) Nature (London) 299, 529-534.

6. Heilman, C. A., Engel, L., Lowy, D. R. & Howley, P. M.(1982) Virology 119, 22-34.

7. Amtmann, E. & Sauer, G. (1982) J. Virol. 43, 59-66.8. Engel, L. W., Heilman, C. A. & Howley, P. M. (1983) J. Virol

47, 516-528.9. Sarver, N., Rabson, M. S., Yang, Y.-C., Byrne, J. C. & How-

ley, P. M. (1984) J. Virol. 52, 377-388.10. Lusky, M., Berg, L., Weiher, H. & Botchan, M. (1983) Mol.

Cell. Biol. 3, 1108-1122.11. Nakabayashi, Y., Chattopadhyay, S. K. & Lowy, D. R. (1983)

Proc. Natl. Acad. Sci. USA 80, 5832-5836.12. Lusky, M. & Botchan, M. R. (1984) Cell 36, 391-401.13. Hanahan, D. (1983) J. Mol. Biol. 166, 557-580.14. Graham, F. L. & van der Eb, A. J. (1973) Virology 52, 456-

467.15. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter,

W. J. (1979) Biochemistry 18, 5294-5299.16. Aviv, H. & Leder, P. (1972) Proc. Natl. Acad. Sci. USA 75,

1274-1278.17. Okayama, H. & Berg, P. (1983) Mol. Cell. Biol. 3, 280-289.18. Grunstein, M. & Hogness, D. S. (1975) Proc. Natl. Acad. Sci.

USA 72, 3961-3965.19. Rigby, P., Rhodes, D., Dieckmann, M. & Berg, P. (1977) J.

Mol. Biol. 113, 237-251.20. Maxam, A. M. & Gilbert, W. (1977) Proc. Natl. Acad. Sci.

USA 74, 560-564.21. Southern, E. M. (1975) J. Mol. Biol. 98, 503-517.22. Seif, I., Khoury, G. & Dhar, R. (1979) Nucleic Acids Res. 6,

3387-3398.23. Ahola, H., Stenlund, A., Moreno-Lopez, J. & Pettersson, U.

(1983) Nucleic Acids Res. 11, 2639-2650.24. Benoist, C. & Chambon, P. (1981) Nature (London) 290, 304-

310.25. Kozak, M. (1981) Nucleic Acids Res. 9, 5233-5252.26. Fitzgerald, M. & Shenk, T. (1981) Cell 24, 251-260.27. Schiller, J. T., Vass, W. C. & Lowy, D. R. (1984) Proc. Natl.

Acad. Sci. USA 81, 7880-7884.28. Danos, O., Engel, L. W., Chen, E. Y., Yaniv, M. & Howley,

P. M. (1983) J. Virol. 46, 557-566.29. Schwartz, E., Durst, M., Demankowski, C., Lattermann, O.,

Zech, R., Wolfsperger, E., Suhai, S. & zur Hausen, H. (1983)EMBO J. 2, 2341-2348.

30. Danos, 0. & Yaniv, M. (1984) in Oncogenes and Viral Genes,eds. Vande Woude, G., Levine, A. J., Topp, W. C. & Watson,J. D. (Cold Spring Harbor Laboratory, Cold Spring Harbor,NY), pp. 291-294.

31. Bouck, N., Beales, N., Shenk, T., Berg, P. & diMayorca, G.(1978) Proc. Natl. Acad. Sci. USA 75, 2473-2477.

32. Martin, R. G., Setlow, V. P., Edwards, C. A. F. & Vembu, D.(1979) Cell 17, 635-643.

1034 Biochemistry: Yang et aL